A Glossary on rapid estimates harmonisation of terminology
- Slides: 18
A Glossary on rapid estimates: harmonisation of terminology to enhance common understanding Rosa Ruggeri Cannata – Eurostat Gian Luigi Mazzi NTTS – Bruxelles 14 -16 March 2017
Glossary on Rapid Estimates Introduction • Lack of common terminology among countries and institutions when talking about rapid estimates • Same terminology used in very different contexts • Communication and common understanding issues • Need for a common vocabulary for various types of rapid estimates • Generally agreed • Based on a transparent and easily understandable logical framework • Eurostat leading the preparation of the glossary on rapid estimates 2
Examples of different terminologies • • Early estimate Flash estimate Nowcasting Rapid estimate 1 st, 2 nd estimate Advanced estimate Preliminary estimate Just a question of when are data released? Eurostat
Issues of interest Eurostat
Structure of the glossary I • Glossary built up around 4 main questions • Each question related to one or more axes of a theoretical hypercube • Each axis has a number of modalities 5
Structure of the glossary II Main questions: • Who? Who makes the evaluation (1 axis). • What? What is evaluated (2 axes). • How? How is the evaluation done (3 axes). • When? When is the evaluation done (2 axes). 6
Structure of the glossary III Who makes the evaluation (1 axis) Axis 1: The uniqueness of an official release vs. the potential multiplicity of evaluations • Producer of rapid estimates may or may not be the same as the producer of regular releases of a given indicator • Possible modalities • Statistical offices or members of the statistical system • Other governmental institutions • Private institutions 7
Structure of the glossary IV What is evaluated (2 axes) Axis 2. The target variable • Possible modalities • • Hard data Soft data Financial data Unconventional data 8
Structure of the glossary V What is evaluated (2 axes) Axis 3. Some revisions in the estimate • Theoretically speaking only data which is characterised by revisions can be the object of flash estimates or nowcasting but also data not subject to revisions can be forecasted • Possible modalities • Data subsequently revised • Data is not revised 9
Structure of the glossary VI How is the evaluation done (3 axes) Axis 4. The adherence to the regular production process • Possible modalities • Fully adherent to the regular production process • Partially adherent to the regular production process • Different than the regular production process 10
Structure of the glossary VII How is the evaluation done (3 axes) Axis 5. Information set • When estimating the target variable the information set on which the estimation is based may or may not include the totality of the information • In case of an incomplete coverage, statistical modelling used to fill the gaps • Defining a minimum acceptable coverage for each estimate • Possible modalities • Availability of the full information set for the period under estimation • Incomplete observation set for the period under estimation • Some variables could be observed only partially • No available information for the period under estimation
Structure of the glossary VIII How is the evaluation done (3 axes) Axis 6. Model/versus parameter uncertainty • Models used for rapid estimates differing for several reasons • Known/unknown data • Techniques implying parameters estimation (uncertainty) vs. Simple smoothing or adjustment techniques • Possible modalities • Statistical models • Econometric models 12
Structure of the glossary IX When is the evaluation done (2 axes) Axis 7. A proper reporting time • In defining rapid estimates the point in time at which they are produced is an essential discriminant • Obviously the frequency of the target variable influences the interpretation of various estimates • Possible modalities • Estimates produced before the reference period • Estimates produced during the reference period • Estimates produced after the end of the reference period, but not later than T+1/2 • … 13
Structure of the glossary X When is the evaluation done (2 axes) Axis 8: Stock and flow data/collecting and reference period • When data are collected and how they are defined also affect the interpretation of various estimates • A regular estimate for a flow variable cannot be produced before the end of the period while for a stock variable recorded at a given day or week of the reference period this would be possible • Possible modalities • Flow • Stock 14
Examples - Nowcasting • Produced by a statistical authority or an institution outside a statistical system • Target variable: hard data • Taking place for the reference period T during the period T itself or right at the end • Making use of all available information becoming available between T-1 and T until the estimation time • Using statistical and/or econometric models different from the regular production process • Hard, soft, financial, unconventional data 15
Examples – Flash estimates • Produced by statistical institutions in charge of the regular production of the concerned indicator • Target variable: hard data • Using an incomplete set of information exploiting as much as possible all available hard data • Soft data can be used to fill some gaps • Using as much as possible the same methodology as for regular estimates • Statistical techniques to deal with incomplete information set • Released as timely as possible after the end of the reference period • Ideally not later than T+1/2 16
Conclusions • Setting up a glossary on the estimation of key economic indicators is not an easy task • A shared operational terminology is necessary to have a common understanding of "what is what" • We have defined the key elements (dimensions) to be taken into consideration • The glossary will be publically available on Eurostat website (SE webpages) Eurostat
Thank you for your attention! Eurostat
- Global harmonisation task force
- Benefits harmonisation
- What is salary harmonization
- Harmonisation des temps verbaux
- Building maintenance cost estimates
- The account analysis method estimates cost functions
- Eva estimates that 475 songs
- Job order cost system
- Marquis company estimates that annual manufacturing
- Properties of least squares estimator proof
- Tom murphy physics
- Who global estimates on prevalence of hearing loss 2020
- Dcms economic estimates
- Was sind fermi aufgaben
- Who global estimates on prevalence of hearing loss 2020
- How photosynthesis takes place
- Ap spanish literature and culture reading list
- Pahina ng nilalaman
- Managed futures glossary